The numerology of the sample directory listing in the Windows 95 font selector property sheet page

In Windows 95, it became tradition that if you modified the code that displays the font preview in the Windows 95 command prompt font chooser property sheet page, you also got to tweak the sample date and file sizes to be a value personally significant to you.

By the time Windows 95 shipped, the values in the font preview were as follows:

Directory of C:\WINDOWS
SYSTEM       <DIR>     03-01-95
WIN      COM    22,867 03-01-95
WIN      INI    11,728 03-01-95
WELCOME  EXE    19,539 03-01-95

I don't see the significance to these values, but maybe you can try to find something. Unfortunately, I don't know who the last person was to modify the dialog template, so I can't ask them.

Bonus chatter: When I told this story to one of my colleagues, I received a story in exchange: When this dialog was ported to Windows NT, the developer who did this slipped his name as the last "file" of the the directory listing, which was visible only when you picked the smallest font size. It was discovered and removed as part of the Easter egg purge.

Riffing on Larry, 13 years later: A note on that rogue Exchange 5.0 Easter egg that my program kept triggering by mistake: Triggering the Easter egg by mistake is bad enough, but what made it even worse was that once you triggered it, the session was poisoned: Commands would behave erratically from that point onward. You had to disconnect from the server and reconnect in order to fix the problem. So yeah, bad Easter egg.

Comments (31)
  1. Brian_EE says:

    The file size of WIN.COM contains 867, and the file size of WELCOME.EXE contains 539, which when combined is *close* to 867-5309, which everyone knows is Jenny’s phone number.

    1. Brian_EE says:

      The lead singer, Tommy Heath, went on to become a software engineer. Did he work for Microsoft in 1995?

  2. outadoc says:

    Can you please re-host the image? is blocked by some firewalls.

    The “easter egg purge” is so interesting. I always kind of wondered why there were less and less easter eggs in Microsoft products. Now I know.

    1. Joshua says:

      While I generally disapprove of the Easter egg purge, I have to admit that exchange Easter egg was horrifying.

      The most recent Easter egg I placed was an extra MVC mapping so that if somebody requested the “directory” containing all the API POST methods, they got a line of static text.

      1. Voo says:

        I hope you do have a spec for your Easter eggs and inform your testers about it.

        The alternative would be to ship untested and unreviewed code in a product that apparently hosts a web server. Something that doesn’t inspire confidence, so I hope it’s the former.

        1. Joshua says:

          Easter egg code gets reviewed, tested by engineering. It doesn’t appear in the API specs only because it isn’t contractual.

          This particular egg has a purpose. It permits bisection testing of the results streamer w/o running the the the result stream generator. If a .NET update breaks the way we stream results we can tell.

    2. viila says:

      I on the other hand do approve of the principle behind the Easter Egg Purge, but my solution would be that any easter eggs have to be part of the design spec, instead of excising them out completely.

      1. Antonio Rodríguez says:

        Then they wouldn’t be Easter eggs. By definition, an Easter egg is some piece of hidden functionality. And if you document it, it isn’t hidden anymore.

        1. Brian says:

          Most users don’t read the documentation. Documenting your easter eggs doesn’t ruin the surprise.

          1. Ian Yates says:

            Document for testers at least… Leave it out of end-user docs.

            It’s like the parents hiding the eggs for the kids – the adults are all in cahoots but the kids still get the fun

  3. Scott H. says:

    Had to check the NT directory listing in my NT 3.51 VM, and there it is. Interestingly I can’t find any mention of this on any of the easter egg database sites, so it may have been previously unknown.

    1. Heh I started my NT 4 VM to check it, but it wasn’t there – probably the EE purge happened just after 3.51.

  4. Pietro Gagliardi (andlabs) says:


    If both myself and Brian_EE are right, I’m going to guess the 22,000 part comes from double the 11,000 part.

  5. Danny says:

    From the link of the purge:
    “When I was discussing this topic with Raymond Chen, he pointed out that his real-world IMAP client hit this bug – and he was more than slightly upset at us for it.”
    So this was Ray using his legendary thermonuclear skills on Larry, huh? That would’ve been a sight to behold :P

  6. BZ says:

    Interesting, I wonder if it’s still being maintained. My version of Windows 8 has all the files updated 10-01-99. Does that mean it hasn’t changed since Windows 2000? Also, the actual output of dir, at least for me, shows results in the format “date time size name”, but the example is still in the order shown above (though with different files and directories)

    1. ChDF T says:

      For future reference, the full text in Windows 7 is:
      C:\Windows> dir
      SYSTEM <DIR> 10-01-99 5:00a
      SYSTEM32 <DIR> 10-01-99 5:00a
      README TXT 26926 10-01-99 5:00a
      WINDOWS BMP 46080 10-01-99 5:00a
      NOTEPAD EXE 337232 10-01-99 5:00a
      CLOCK AVI 39594 10-01-99 5:00p
      WIN INI 7005 10-01-99 5:00a

      It is noteworthy that:
      – The text below README and everything to the right of DIR is not visible in the actual property page, therefore the am/pm change for CLOCK.AVI and the file sizes are not displayed.
      – The text does not seem to be localized (the above text was captured on a system localized to de-de).
      – The files README.TXT, WINDOWS.BMP, CLOCK.AVI did not exist on the test system; the files NOTEPAD.EXE (notepad.exe) and WIN.INI (win.ini) use different casing, which is reflected when actually executing dir.
      – The files in the sample output are not sorted alphabetically by name, like the actual dir does by default in the tested OS.

  7. B says:

    I would guess they are birthdays in day, month year format.
    22/8/67 (would have been 27 years old)
    11/7/28 (would have been 66 years old)
    19/5/39 (would have been 56 years old)

    1. viila says:

      That was my first instinct, but DD/M/YY is wrong for US, and someone else pointed out 11728*2==23456 which seems much too exact to be coincidence, so I don’t think they’re dates.

      1. morlamweb says:

        @viila: DD/M/YY is not in common use in the US, but I wouldn’t call it “wrong”. It’s a date format like any other. Using a date format which is not common in the US, and then formatting it as a number, would be right in line with a hidden message.

  8. Richard Wells says:

    Provides a good range of characters: narrow, wide and even a lower case “y” for the descender. The strange element is the number “4” is omitted from the sample though showing the differences between the “1” versus “I” and the “O” versus “0” are more relevant.

  9. BJC says:

    The files sizes could be translated to dates, in UK format (DD-M-YY). Perhaps the first is the developer birth date (22-Aug-67) born to parents with birth dates of 11-Jul-28 and 19-May-39. More likely this is just the human ability to pick out a pattern where none exists – at least for this suggestion.

  10. Pietro Gagliardi (andlabs) says:

    Also just to entertain the idea, here’s the upside down seven-segment reading:

    22,867 -> LgBZZ
    11,728 -> BZLII
    19,539 -> GESGI

    Probably doesn’t mean anything…?

  11. Teijo says:

    22867 = hex 5953, in ASCII “YS”
    11728 = hex 2DD0 ->”-Д (whatever that means, some kind of smiley perhaps?)
    19539 = hex 4C53 -> “LS”

    At least the first and last number appear to be encoded initials.

    1. R P (MSFT) says:

      0xD0 in CP850 is a lowercase eth (ð), not that that helps.

  12. Drak says:

    I recall placing an ‘easter egg’ (more of a reference to a certain adult oriented adventure game) in a bit of code when we first used SOAP. I put a comment in just before the call to a SOAP service saying .

    A client discovered this and called our help desk to inquire as to what it meant.

  13. DWalker07 says:

    “the Windows 95 command prompt font chooser property sheet page…” That’s a mouthful!

    1. Imagine how long that German translation would be!

      1. Chrissielein says:

        Windows 95 Befehlszeilenschriftartauswahlfensterregisterkarte?

        Rolls right off the tongue…

      2. florian says:

        Die Windows 95-Befehlszeilenkonsolenschriftartenauswahldialogeigenschaftsseite.

        Didn’t manage to count the characters, is it really longer than in English?

        Also: putting hyphens between the nouns in different ways could change the meaning in subtle ways.

        For everyday use, you would probably put it like this:

        Die Eigenschaftsseite des Dialogs zum Auswählen der Schriftart für die Befehlszeilenkonsole.

        Now it’s definitely longer than in English.

        1. keal says:

          @florian: “Also: putting hyphens between the nouns in different ways could change the meaning in subtle ways.”

          In the vinyl record days, with pick ups where either the coils or the magnet were moving, I read one article about some “moving
          coil-pick ups”

          I had never before seen a coil-pick, and certainly not a moving one (or maybe it would be a few ups that were both coil-pick, whatever that is, and moving at the same time), so it took me a couple of seconds to grasp it. The line break before “coil-pick” (which was there) didn’t improve readability.

      3. Brian_EE says:

        die Windows 95-Eingabeaufforderung Schriftart-Auswahl-Eigenschaftenblatt-Seite

Comments are closed.

Skip to main content